Hello, i tried to use GENX320 Event Camera Module for face regonition in this way. created dataset after converting the event-based streams into farmes then croped the farmes with MediaPipe , and trained the model with Edge Impulse , then deployed. the accuray of classification was good but to recognize the face , it is little bit not easy, i know that the event-based cam detece moving objects, so that faces must move to much to get recognized.
How do you advice me to improve this face recognition with GENX320 Event Camera Module.
Hi, the GenX320 wouldn’t really be the camera you’d want to use for face recognition. It can’t see stationary objects… so… not sure how it would be useful detecting a face that’s staring into a camera…
ok , thanks for ur reply. i changed my setup so that the Genx320 will be moving will detecting faces . Now, i’m trying to implement the solution u proposed before on how to get the raw data (x,y,t,p) by changing the firmware but it dose not work. Any advice on this matter.
We are working on a firmware release at the moment which will be out soon which will package everything up nicely with docs and examples and etc. in the IDE.